The investigations described show that it is possible to extract specific substances from V. choleræ by means of hot dilute alcohol. This result cannot be attributed to the presence of bacilli in the extracts as would follow from the statements made by Prausnitz in his criticism of the work of Levaditi. The original extract contains protein and exhibits antigenic properties. From this preparation an almost protein-free product was obtained. It was fully active in the precipitin test but had lost almost completely the antigenic activity. Accordingly the latter preparation belongs to the class of substances described by Zinsser as residue antigens and studied chemically by Avery and Heidelberger. The fact that the immune sera resulting from the injections of the crude extracts acted upon the non-immunizing precipitable substance indicates that in the crude extract there is present an antigenic complex consisting of protein and the specific substance. Regarding its chemical nature it follows from the foregoing that the precipitable but non-immunizing substance is not a protein. On hydrolysis it yielded a considerable quantity of sugar although less than that given for the specific carbohydrates of pneumococci and B. friedlænderi. The product contains nitrogen and phosphorus and on hydrolysis a substance of acid character separates from the solution. Accordingly the substance prepared would appear to have either a rather intricate structure or to be a complex carbohydrate, similar to those described by Avery and Heidelberger, but still containing impurities. This issue can probably be decided by further studies.